Chapter 3B Y - i Upper-Atmosphere Aerosols: Properties and Natural Cycles r " Richard Turco , University of California \ '" Los Angeles, CA ABSTRACT The middle atmosphere is rich in its variety of particulate matter, which ranges from meteoritic debris, to sulfate aerosols, to polar stratospheric ice clouds. Volcanic eruptions strongly perturb the stratospheric sulfate (Junge) layer. High-altitude "noctilucent" ice clouds condense at the summer mesopause. The properties of these particles, including their composi- tion, sizes, and geographical distribution, are discussed, and their global effects, including chemical, radiative, and climatic roles, are reviewed. Polar stratospheric clouds (PSCs) are composed of water and nitric acid in the form of micron-sized ice crystals. These particles cat- alyze reactions of chlorine compounds that '~activate"'otherwisein ert chlorine reservoirs, leading to severe ozone depletions in the southern polar stratosphere during austral spring. PSCs also modify the composition of the polar stratosphere through complex physicochemical processes, including dehydration and denitrification, and the conversion of reactive nitrogen oxides into nitric acid. If water vapor and nitric acid concentrations are enhanced by high-altitude aircraft activity, the frequency, geographical range, and duration of PSCs might increase accordingly, thus enhancing the destruction of the ozone layer (which would be naturally limited in geographical extent by the same factors that confine the ozone hole to high latitudes in winter). The stratospheric sulfate aerosol layer reflects solar radiation and increases the planetary albedo, thereby cooling the surface and possibly altering the climate. Major volcanic eruptions, which increase the sulfate aerosol burden by a factor of 100 or more, may cause significant global climatic anomalies. Sulfate aerosols might also be capable of activating stratospheric chlorine reservoirs on a global scale (unlike PSCs, which represent a localized polar winter phe- nomenon), although existing evidence suggests relatively minor perturbations in chlorine chemistry. Nevertheless, if atmospheric concentrations of chlorine (associated with anthro- pogenic use of chlorofluorocarbons) continue to increase by a factor of two or more in future decades, aircraft emissions of sulfur dioxide and water vapor may take on greater significance. .a4 ..--- --.-- INTRODUCTION Particles and clouds in the stratosphere and mesosphere have been under study for more than 100 years. High-altitude aerosols were the subject of scientific speculation during the 1880s, when the powerful eruption of Krakatoa caused spectacular optical displays worldwide, attracting attention to the upper atmosphere. The presence of a permanent tenuous particle layer in the lower stratosphere was postulated in the 1920s through studies of the twilight glow (1). The first in situ samples of these particles showed they are composed of sulfates, most likely concentrated sulfuric acid (2-5). Subsequent research was spurred by the realization that strato- spheric particles can influence the surface climate of Earth by modifying atmospheric radiation (6). Such aerosols can also affect the trace composition of the atmosphere, ozone concentra- tions, and the electrical properties of air (7). Stratospheric particulates have been sampled by balloon ascents and high-altitude aircraft to determine their properties directly (8). The aerosols have also been observed remotely from the ground and from satellites using both active (lidar) and passive (solar occultation) techniques (remote sensing instruments have been canied on aircraft platforms as well) (9,lO). In connec- tion with the experimental work, models have been developed to test theories of particle forma- tion and evolution, to guide measurement strategies, to provide a means of integrating labora- tory and field data, and to apply the diverse scientific knowledge gained to answer practical questions related to issues of global changes in climate, depletion of the ozone layer, and related environmental problems (1 1). In the following sections, primarily stratospheric, but also mesospheric, particles are described, and their global effects are discussed. Figure 1 illustrates many of the species of 65 kK~EDlffPi AGE BLANK NOT FtLMED aerosols that have been identified in the upper atmosphere. The diagram provides information on the size dispersion and concentration of these diverse particulates (12). Table 1 provides a summary of the key characteristics of stratospheric aerosols. Table I!. Characteristics of .Stratospheric Aerosols Particle Sulfate Type-I Type-I1 Meteoric Rocket type aerosol PSC PSC dust exhaust Physical Liquid or Solid nitric Solid crystal, Solid granular Solid spheres state slurry with acid trihydrate, hexagonal or irregular or or irregular crystals solid solutions cubic basis spherical surface ablated debris 0.01 - 0.5, Arnb. 1 - 100, Micro- 0.01 - 10, Volc. meteorites 0.01 -0.1, smoke Principali SiO2, Fe, composition Ni, Mg; C HCI HN03, S042- HCI Physical Dust inclusions, Equidimensional Elongated =-re@ Homogeneous cbcreristics in solution crystalline or crystals with mineral grains, composition; droplets poly crystalline grain defects smooth spheres structure STRATOSPHERIC SULFATE AEROSOLS The presence of trace amounts of sulfur-bearing gases in the stratosphere favors the fomatio~no f sulfuric acid aerosols (13-18). In particular, carbonyl sulfide in the background atmosphere is largely responsible (19) for the tenuous, ubiquitous ambient sulfate haze observed in the stratosphere (20-21). The general properties of these aerosols are summarized in Table 2. The processes that control this haze also influence the formation and evolution of volcanically induced aerosols. Accordingly, investigations of the ambient stratospheric aerosol Payer prcbvide insights into the behavior of volcanic eruption clouds above the tropopause. Backgr~ound Aerosols kt is fairly well established now that the ambient stratospheric sulfate layer is formed as a result of the chemical transformation of sulfur-bearing gases (carbonyl sulfide, OCS, and sulfur &oxide, SO2) transported into the stratosphere from the troposphere or injected there by major volcanic eruptions (19). Chemical reactions of these precursor sulfur gases lead to the produc- 10 10 a 10-1 100 10 10 I 03 Particle Radius (Clm) Figure 1. Spectrum of particles in the Earth's middle atmosphere. Shown are the approximate size distributions for particles with different origins. The total (number) concentration of each type of particle is roughly indicated by the peak value on the vertical scale. Some aerosols are highly variable in concentration and properties, particularly the "ash" component of volcanic eruption clouds; typically, these particles will fall out of the stratosphere within a few months following an eruption. tion of condensible sulfur compounds - primarily H2S04 [detected in situ by Arnold et al., (22)l. The sulfur conversion process is dominated by the reaction sequence: (23) SO2 +OH + M ----> HS03 + M (I> HS03 + 0 2 ----> SO3 + H02 (2) SO3 + H20 ----> H2S04 (3) "9"alSle 2. Properties of Stratospheric Sulfate Aerosols Composition H2SO&20 (-70%/30%); traces of sulfates, nitrates, nitryls, chlorides &gin OCS, CS2 (also volcanic sulfur emissions); tropospheric sulfides; SO2 photochemical oxidation to H2S04v ia OH; high-altitude aircraft S@ emissions (contribution unknown) Liquid sphe-re s, perhaps slurry and some solids; < 1 ppbm; -1- 10 cm-3 0.05 pm radius Dis~bution Global, 12 to 30 km altitude; latitudinal and seasonal variations - - Mass Budget 0.1 Tg-S/yr (background); 1-100 Tg-S (volcanic event) - Residence Time 1-2 yr (average, based on radioactive tracer studies) Effects Shortwave radiation scattering (z< 0.01); longwave absorption/emission (z<< 0.01); heterogeneous chemical conversion of NOx to HNO3 Influences Natural and anthropogenic OCS sources; volcanic activity; stratospheric dynamics Variable over 2 orders of magnitude following major volcanic e-r uptions (e.g., El Chichon). Possible long-term increase of 6%/yr by mass Reactions (2) and (3) are so rapid that sulfur radicals (e.g., HS03 ) never achieve a significant concentration, and thus do not play a role in the chemical evolution of the aerosol cloud (except as an intermediary sulfur species). Although reaction (3) may require the presence of surfaces to occur rapidly, it does not limit the overall oxidation rate of sulfur dioxide. Importantly, reac- tion sequence (1) and (2) does not consume odd-hydrogen, HO,, as would the competing pro- cess consisting of reactions (1) and (4), Hence, HOx oxidizes SO2 catalytically, and the rate of SO2 conversion can remain high thoughout the evolution of a volcanic cloud provided there is sufficient recycling of H02 to OH. 'The properties of the sulfate particle layer are strongly influenced by microphysical pro- cesses, including heterogeneous nucleation, growth by condensation, evaporation of volatiles, coagulation, and gravitational sedimentation (24). Some of the complex physicochemical inter- ac~onrse sponsible for the formation of the stratospheric aerosol layer are illustrated schemati- cally in Figure 2; the potential contributions of naturally occuning meteoritic particles and ion clusters to sulfate aerosol formation are also indicated. Exhaust particles from Space Shuttle or high-dtitude ahraft operations would enhance the condensation nuclei abundances. Detailed discussions of microphysical processes can be found in several papers, reviews, and books (25,26). rn I I GALACTIC METEOR ABLATION J 0 S P H E R METEORITES E MICROMETEORITES COAGULATION T R 0 NUCLEATIO* P SULFUR VAPORS tso,. m. c q P SEDIMENTATION H DIFFUSION E WASHOUT PROOUCTtON R EMISSION Figure 2. Diagram depicting the physical and chemical processes that affect stratospheric aerosols. Meteoritic debris, positive and negative ions, and tropospheric Aitken nuclei can serve as nucleation sources for sulfate aerosols and ice clouds. These particles can grow, evaporate, coagulrjlle, and fall vertically. The particles are also carried in stratospheric winds and diffuse under the influence of small scale turbulent eddies. Volcanic Aerosols Table 3 summarizes the general properties of the aerosols generated by a modest-to-large volcanic eruptions (that is, one in which the eruption column penetrates the tropopause and deposits substantial quantities of gases and particles in the stratosphere) (27-32). The general microphysical development and properties of volcanic eruption plumes are exemplified by the behavior of the El Chichon eruption cloud (33,34). The primary c~mptiono f El Chichon occurred on April 4,1982; solid debris and gases were injected to altitudes of about 30 km over the Yucatan Peninsula (35). Many observations of the clouds were made:, and these data have been compared to model simulations that include the relevant physical and chemical processes (36,37). It should be noted that high-altitude sampling of the El Chicholn volcanic clouds was very limited. Data collected by the Solar Mesosphere Explorer satellite suggest that the post-eruption aerosols reached altitudes of 40 km (38,39). Lidar measurements, on. the other hand, indicated that the El Chichon particles remained below about 30 km for several. months after the eruption. Table 3. Properties of Volcanic Aerosols Composidon Silicates; H2S04/I120( -70%/30%); traces of sulfates, nitrates, chlorides, etc. Ol;i& Tenrestrial material, gaseous SO;! with chemical oxidation to &SO4 via OH I>Hopfies Liquid spheres; solid mineral particles dominant the fxst month; - - 100-1000 ppbm; (highly variable); -1-10 cm-3; 0.3 pm radius Dis~bntion Regional (days); zonal (weeks); hemispheric (months); global (year); 12-35 km altitude - Mass Budget Per event: S@ (- 1-100 Tg-S); N20 ( 10-1 000 Tg); HCl (- 0.01- 10 Tg); mineral ash (> 100-10,000 Tg) - Residence Time 1-3 yr (average, based on radioactive tracer and aerosol decay studies) Effects Shortwave scattering of sunlight leads to surface cooling; Longwave absorption warms the stratosphere; Injection of H20, HCl, etc., possibly alters composition; Enhanced heterogeneous reactions on sulfate aerosol surfaces; Possible ozone perturbations; Stratospheric stability / tropospheric dynamics affected; Nuclei for upper tropospheric cirrus Inf uences Geophysical; no anthropogenic influences; geological setting determines effects; impact on ozone may be affected by chlorine levels Trrenh Random significant eruptions are -20 years on average; major eruptions are -100 years apart 'The size distributions of volcanic aerosols (shown in Figure 3 for an El Chichon simu- la~one)x hibit a ~-modaslt ructure that evolves with time. The principal size modes are: a nucleation mode, which is most prominent at early times and at sizes near 0.01 pm; a sulfate accumulation mode, which evolves initially from the nucleation mode (by coagulation and con- densation) and increases in size to about 0.3 pm after 1 year; and a large-particle "ash" mode (sf solid mineral and salt particles) that settles out of the layer in 1 or 2 months. A primary feature of the volcanic aerosol size distribution after several months is a greatly enhanced sulfate ac.ccumu:Laeonm ode. The increased aerosol size is caused by accelerated growth in the presence of enhanced sulfuric acid vapor concentrations that are maintained by continuing SO2 chemical conversion. ]Figure 4 illustrates the evolution of volcanic aerosol optical depths at mid-visible wave- lengths, associated with the scattering of light by the sulfuric acid droplets. The calculations cornspond to the sulfur injection scenarios used by Pinto et al. (40); i.e., SO2 mass injections of 10 Tg (1 Tg = 1 x 1012 g = 106 metric tons) (i.e., like El Chichon), and 100 and 200 Tg (pssibly similar to Tambora, 18 15). At early times, and over limited geographical regions, the optical depths can exceed a value of 2. I-Iowever, after 1 year of evolution by growth, coagula- tion, and fallout, the average optical depth for even the largest SO2 injection has fallen to about 0.5. These results suggest that nonlinear physicaVchemical interactions occurring in volcmic eruption clouds severely limit the aerosol optical depth that can be maintained over a. pefid of several years (40) (i.e., the time span required to induce substantial long-term climatic impaclts, see also section on Radiation and Climate Effects). Conversely, the efficiency for producing radiative effects per unit mass of sulfur injected is greatest for smaller injections. Accsrdjingly, high-altitude aircraft emissions of SO2 hold the potential for creating significant global-scale radiative effects. Figure 3. Evolution of the aerosol size distribution at 20 km in the simulated El Chichon eruption cloud. Size distributions are shown at various times, and are compared to the ambient size diisthbution (36). 100 MT SO2, 5 x 1 o7 krn2 100 MT SO2, 1 x 1 o8 krn2 GLOBAL DISPERSION 0 1 2 3 6 9 12 15 18 2 1 24 TIME (mo) Figure 4. Volcanic aerosol optical depths (zenithal) versus time for the volcanic cloud simulations discussed by Pinto et al. (40). Indicated on the figure are the times required for the cloud to disperse over a hemisphere, or over the globe (40). The sulfate aerosol mode radius peaks earlier with larger SO2 injections and remains elevateti throughout the history of the eruption cloud. In the simulations shown, the mode radius grows as large as 0.7 pm, which greatly exceeds the ambient sulfate mode radius, r-0.05 para. ]In the instance of very large eruptions, the mode radius returns to its ambient value only after a period of several years. The mode radius is important in determining the rate at which sulfate is removed from the stratosphere. The sulfate mass flux from the stratosphere is propodonal to the fallspeed, v, of the aerosols multiplied by their mass, m. Since m = r3 and v oc r1+2 in the regime of interest, the mass loss rate is, m = 4-5 The optical depth per unit mass of aerosol varies roughly as l/r in the size range of interest. Hence, the decrease in optical depth in Figure 4 can be seen to have two causes: (1) the growth in particle size that reduces the oprical efficiency; and (2) a rapid decrease in the the total sulfate mass caused by sedimentation (after a month or so). These nonlinear interactions greatly limit the potential climatic effects of explosive volcanism. POLAR STRATOSPHERIC CLOUDS (PSCs) The properties of stratospheric clouds in polar regions (PSCs), (41) have been defined by a de'cade of satellite observations (see Table 4) (42-45). Based on optical and physical evi- dence, PSCs fall into two broad categories, which are referred to here as Type I and Type I1 PSCs. Type I PSCs consist of an aerosol haze of micron-sized nitric acid ice particles com- posed of II[N03 and H20 [in roughly a 50150 mixture by weight, similar to the trihydrate HN03.3H20] (46,47). Type I1 PSCs are apparently composed of water-ice crystals (48). Some s.catistica1p roperties of PSCs derived from satellite observations are summarized in Figure 5 (45). 72 Table 4. Properties of Polar Stratospheric Clouds Composition Type I: HN@m20 (-50%/50%); Type-11: water ice; possibly traces of HCl, HN03, etc. Type I: nucleation on sulfate, Te195 K, Type II: ice nucleation sn Typ I, Tc189 K Properties Type I: -1-10 ppbm; <1 cm-3; 0.3 pm; Type 11: -1 ppmm; <el em-3; 3 pm radius; solid crystalline structures Distribution Polar winter stratospheres (>60° latitude); 14-24 km altitude; winter and early spring; S. H., June-October, widespread Type I and IH; N. H., December-March, sporadic I, occasional I1 - - Mass Budget Type I: 1-10 ppbm HN03 per winter season; Type 11: 1-5 ppmm H20 per winter - Residence Time Type I: 1 day, to weeks (temperature control, to sedimentation control); - Type 11: hours (condensation/sedimentation/evaporation con~ol) Effects Activation of chlorine reservoirs (Type I and II); conversion of .NOxi nto IIN03 (Type I); dehydration of the polar stratosphere (Type D); denitrification of the polar vortex (Type I and 11) Influences Stratospheric polar meteorology; tropospherically driven wave events; vortex stability and temperature; springtime warming events; pssible role of changes in C&, H20 and NOy Trends Tied to trends in polar meteorology, especially temperature The Type I PSCs are considered the most common form, accounting for perhaps 80 to 90% of all cloud sightings. These PSCs exhibit an onset at temperatures near 195 K, whereas the more massive Type I1 PSCs appear to condense at colder temperatures (el87 K) consistent with the measured frost point of water vapor in the polar stratosphere (48,49). It should be expected, therefore, that Type I haze would predominate the totality of cloud observatiolns in the earliest part of the Antarctic winter season, and that the frequency of Type I1 clouds would increase with the progression of winter and cooling of upper air layers. On the other hand, the observed dehydration and denitrification of the Antarctic winter stratosphere would, over the course of time, reduce the frequency of cloud formation at specific temperature thresholds (50). In the late winter and early spring, predictions and observations indicate that PSCs will (dissipate abruptly when the upper strato-sphere warms (5 1). Most likely, the nuclei for nitric acid ice deposition are the background sulfilfic acid aerosols (46). Observational evidence on the extent to which sulfate particles are nucleated in PSCs is mixed. Aircraft measurements taken in the Antarctic ozone hole in September 1987 suggest that many, if not most, of the sulfate particles may be activated into nitric acid haze par- ticles (52,53). On the other hand, balloon-borne aerosol measurements taken at M[cMurdo Station during the same period, and more recent data from the Arctic winter stratosphere, indi- cate that, although layers of -1 micron-size haze particles are frequently present, on occ;asion the